Multidimensional Workaholism Scale Validated Across Four Core Dimensions for Brazilian Workers

A comprehensive psychometric validation study published in Frontiers in Public Health has culturally adapted and validated the Multidimensional Workaholism Scale for the Brazilian workforce, establishing the Escala Multidimensional de Workaholism as a psychometrically sound research instrument. Conducted by an interdisciplinary research team affiliated with Brazilian and Portuguese universities, the methodological investigation demonstrated that the sixteen-item adapted instrument reliably captures four distinct yet interrelated operational dimensions: motivational, cognitive, emotional, and behavioral. Utilizing an extensive cross-sectional sample of seven hundred twenty actively employed workers representing diverse economic sectors and regional areas across Brazil, the researchers confirmed the reflective structure of the scale through partial least squares structural equation modeling. The evaluation verified satisfactory internal consistency reliability, robust convergent validity based on average variance extracted metrics, and empirical discriminant validity across all latent constructs. Furthermore, rigorous multigroup analysis through the measurement invariance of composite models procedure confirmed full measurement invariance between male and female workers, demonstrating that the operational structure of the instrument functions equivalently across sex groups.

Workaholism has increasingly been recognized in modern organizational and psychological research as a prominent psychosocial occupational risk characterized by excessive and compulsive work behaviors that extend far beyond formal contractual requirements. Modern employment environments, marked by widespread digital connectivity, intensified performance demands, and the erosion of boundaries separating occupational obligations from personal life, have heightened employee exposure to psychosocial risks that undermine well-being. Empirical literature consistently links workaholic patterns with severe health consequences, including emotional exhaustion, sleep disturbances, psychological distress, anxiety, chronic burnout, work-family conflict, impaired physiological recovery, and reduced overall quality of life. Despite expanding research into these occupational challenges, the systematic investigation of workaholism has long been hindered by persistent conceptual ambiguity and measurement heterogeneity. The construct frequently overlaps with adaptive forms of heavy work investment, such as legitimate job involvement, healthy work engagement, organizational commitment, and intrinsic motivation, as well as distinct personality dispositions including perfectionism and negative affectivity. Historically, conflicting measurement tools have generated contradictory findings and restricted cross-study comparability, highlighting the necessity for culturally adapted, theoretically grounded evaluation tools.

To overcome the limitations of legacy assessment tools, contemporary research has sought to establish clear operational boundaries separating pathological work investment from productive engagement. Early psychometric instruments, such as the Work Addiction Risk Test, the Workaholism Battery, and the Dutch Workaholism Scale, contributed substantially to initial occupational research but frequently encountered structural instability, conceptual overlap, and inconsistent factorial solutions across different cultural adaptations. In contrast, the Bergen Work Addiction Scale introduced a specialized framework rooted strictly in core addiction symptoms such as salience, tolerance, mood modification, relapse, withdrawal, conflict, and problems; however, its unidimensional architecture precludes the differentiated assessment of distinct psychological components. Addressing this structural gap, the Multidimensional Workaholism Scale developed by Clark and colleagues conceptualized workaholism as four interrelated but distinct dimensions. Under this framework, the motivational dimension reflects an internal and compulsive compulsion to work, the cognitive dimension captures persistent and intrusive work-related thoughts during off-work hours, the emotional dimension encompasses negative affective states experienced when work activities are obstructed, and the behavioral dimension measures overt engagement in excessive work activities beyond normal expectations.

The cultural adaptation and empirical validation of the Brazilian version followed standardized international guidelines for adapting self-report instruments across six systematic stages. The initial phase involved independent forward translations of the original English items into Brazilian Portuguese by two independent bilingual translators, one familiar with occupational psychology and the other uninformed regarding the specific construct. Consensus synthesis reconciled discrepancies between the two translations, followed by blind back-translation into English by two native English speakers fluent in Portuguese. An expert committee comprising five specialists with recognized expertise in psychological assessment, research methodology, and organizational behavior systematically evaluated the semantic, idiomatic, experiential, and conceptual equivalence of the sixteen candidate items. The expert committee scrutinized each item according to clarity, comprehension, pertinence, and theoretical relevance using a five-point Likert scale. Content validity was quantitatively quantified using Lawshe’s Content Validity Ratio, where individual items were assessed against a benchmark of essentiality to ensure that the translated indicators rigorously mirrored the theoretical definitions established in the original scale framework.

The quantitative content validity assessment revealed Content Validity Ratio scores ranging from 0.20 to 1.00 across the sixteen proposed indicators, highlighting varying levels of expert agreement and pointing directly to specific items requiring targeted refinement. While the cognitive dimension exhibited high expert concordance, with three of its four indicators achieving perfect scores of 1.00, the motivational dimension exhibited the greatest need for linguistic revision. Items MOT.03 and MOT.04 yielded Content Validity Ratio values of 0.20 for pertinence and comprehension, while MOT.02 demonstrated potential ambiguity and conceptual overlap with emotional indicators like EMO.03. To resolve these concerns, a sixth independent specialist in linguistics was consulted to refine the Portuguese phrasing, eliminate subtle ambiguities, and sharpen discriminant clarity while preserving the original theoretical core. Following these refinements, a pre-final draft was administered to a pilot sample of one hundred seven Brazilian workers to evaluate clarity and ease of completion. The pilot phase confirmed overall comprehensibility without major interpretive difficulties, preparing the standardized instrument, officially designated as the Escala Multidimensional de Workaholism, for extensive nationwide psychometric testing.

The main psychometric evaluation examined responses from seven hundred twenty Brazilian workers recruited through digital outreach between December 2023 and February 2024, following the exclusion of four cases exhibiting response bias patterns. The final sample consisted of actively employed Brazilian adults aged eighteen to seventy-five years, with a mean age of 40.4 years and standard deviation of 12.3 years, of whom 58.8 percent identified as female. Geographically, respondents were distributed across all five national regions, although residents of the Southern region predominated at 78.5 percent, followed by the Southeast at 7.5 percent, Northeast at 6.4 percent, North at 4.7 percent, and Central-West at 2.9 percent. The participants represented an educated and economically stable cohort: 41.1 percent held doctoral degrees, 18.1 percent had completed undergraduate education, and 37.2 percent reported monthly earnings exceeding ten thousand Brazilian reais. In terms of employment sectors, 59.4 percent were employed in the public sector, 27.1 percent in the private sector, and 8.9 percent were self-employed, with 33.1 percent occupying managerial positions. Work schedule indicators showed intensive involvement: 45.8 percent had contractual forty-hour workloads, 38.3 percent routinely logged forty-one to fifty total weekly hours, and 45.3 percent worked weekends frequently or always.

Preliminary exploratory factor analyses were conducted on the pilot dataset of one hundred seven workers to inspect the internal coherence and behavior of items within each theoretically defined dimension. Principal axis factoring was executed separately for each dimension based on Pearson correlation matrices without rotation, specifying a single-factor extraction in alignment with the original theoretical framework. The sampling adequacy was confirmed by Kaiser-Meyer-Olkin metrics ranging from 0.728 for behavioral indicators to 0.851 for cognitive indicators, while Bartlett’s test of sphericity attained statistical significance across all four dimensions with probability values below 0.001. The proportion of total variance explained by the single extracted factor was 76.57 percent for the cognitive dimension, 72.45 percent for the emotional dimension, 59.53 percent for the motivational dimension, and 53.57 percent for the behavioral dimension. Within-dimension exploratory factor loadings ranged between 0.660 and 0.910 across all sixteen items, and communality estimates confirmed satisfactory shared variance, demonstrating that the individual translated indicators exhibited satisfactory structural cohesion prior to testing the full measurement model in the primary sample.

Confirmatory evaluation of the reflective measurement model was performed on the full sample of seven hundred twenty respondents using partial least squares structural equation modeling with five thousand bootstrap resamples. Global model fit indices demonstrated excellent correspondence with empirical data, evidenced by a standardized root mean square residual of 0.056, well below the established 0.08 benchmark, and a normed fit index of 0.906. Bootstrap validation further confirmed model adequacy, as the observed standardized residual and both unweighted least squares distance and geodesic distance remained strictly beneath their ninety-fifth and ninety-ninth percentile thresholds. All indicator factor loadings exceeded the preferred 0.70 criterion, ranging from 0.782 to 0.906, with the sole exception of MOT.02 at 0.678, which was retained due to its essential theoretical contribution to motivational content coverage. Internal consistency reliability was confirmed across all dimensions: Cronbach’s alpha values ranged from 0.824 to 0.908, Dijkstra-Henseler’s rho_A coefficients ranged from 0.847 to 0.910, and composite reliability coefficients ranged from 0.883 to 0.936. Convergent validity was established by average variance extracted estimates ranging from 0.656 to 0.784, proving that each latent construct accounted for more than half of its indicator variance.

Discriminant validity was established through both the traditional Fornell-Larcker criterion and modern heterotrait-monotrait ratio analysis. For every dimension, the square root of the average variance extracted, which varied from 0.810 for motivation to 0.886 for cognition, exceeded all bivariate correlations between latent constructs, where the highest observed inter-construct correlation was between the cognitive and motivational dimensions at 0.673. In parallel, all ninety-five percent bootstrap confidence intervals for heterotrait-monotrait ratios fell below 0.85, with the upper confidence limit peaking at 0.811 for the cognitive-motivational dimension pair and no interval containing unity. Evaluation of the structural paths linking the four dimensions revealed variance inflation factors between 1.000 and 2.029, verifying that multicollinearity did not distort structural estimations. Path effect sizes ranged from 0.061 to 0.502, with the cognitive construct exerting the strongest structural associations upon emotional and behavioral dimensions. The endogenous constructs yielded significant coefficients of determination ranging from 0.334 to 0.497, alongside predictive relevance values surpassing 0.25, confirming the robust explanatory and predictive capacity of the interconnected multidimensional framework.

A critical contribution of the research was the rigorous verification of measurement invariance across participant sex groups using the three-stage measurement invariance of composite models procedure. Configural invariance was ensured by administering identical item sets, data transformations, and algorithm configurations across the four hundred twenty-three female and two hundred ninety-seven male workers. Compositional invariance was established across all four latent constructs, with original correlation coefficients between 0.999 and 1.000 and permutation probability values exceeding the five percent significance threshold. Subsequent permutation testing revealed no statistically significant differences between male and female composite mean values or composite variances. Further multigroup analyses applying Henseler’s procedure and permutation tests revealed no statistically significant disparities in estimated structural path coefficients between sexes. Despite these strengths, the authors explicitly emphasized important research limitations: the cross-sectional design precludes causal interpretations, convenience sampling concentrated in southern Brazil limits national generalizability, and high educational and public-sector representation may not reflect the broader working population. Crucially, the researchers cautioned that because diagnostic thresholds, sensitivity, specificity, and longitudinal outcomes were not evaluated, the Escala Multidimensional de Workaholism should currently function strictly as an academic research tool rather than an individual screening mechanism or professional diagnostic instrument.

Subject of Research: Psychometric validation of the Multidimensional Workaholism Scale for Brazilian workers

Article Title: Brazilian adaptation and psychometric evaluation of the multidimensional workaholism scale

Article References: Moreira, M. G., Santos, J. V. D., Silva, D. J. C. D., Steil, A. V., Lopes, E. G., Gaspary, J. F. P., & Lopes, L. F. D. (2026). Brazilian adaptation and psychometric evaluation of the multidimensional workaholism scale. Frontiers in Public Health, 14, Article 2011901. https://doi.org/10.3389/fpubh.2026.2011901

Image Credits: AI Generated

DOI: 10.3389/fpubh.2026.2011901

Keywords: workaholism, Multidimensional Workaholism Scale, Escala Multidimensional de Workaholism, psychometrics, scale validation, cross-cultural adaptation, occupational health, measurement invariance, PLS-SEM, convergent validity, discriminant validity, internal consistency reliability, MICOM, psychosocial occupational risks, Brazilian workers, job stress, emotional exhaustion